A5104546839- Reinforcement Learning in Robotics
- Topic Modeling
- Natural Language Processing Techniques
- Adaptive Dynamic Programming Control
- Speech Recognition and Synthesis
- Semantic Web and Ontologies
- Advanced Bandit Algorithms Research
- Complex Systems and Time Series Analysis
- Adversarial Robustness in Machine Learning
- Multi-Agent Systems and Negotiation
- Game Theory and Applications
- Evolutionary Game Theory and Cooperation
- Advanced Memory and Neural Computing
- Research Data Management Practices
- Advanced Neural Network Applications
- Blockchain Technology Applications and Security
- Ethics and Social Impacts of AI
- Sparse and Compressive Sensing Techniques
- Speech and dialogue systems
- Multimodal Machine Learning Applications
- Smart Grid Energy Management
- Sports Analytics and Performance
- Space Satellite Systems and Control
- Human Pose and Action Recognition
- Spacecraft Design and Technology
University of Manchester
2024
King's College London
2024
Shanghai Jiao Tong University
2021-2024
Many artificial intelligence (AI) applications often require multiple intelligent agents to work in a collaborative effort. Efficient learning for intra-agent communication and coordination is an indispensable step towards general AI. In this paper, we take StarCraft combat game as case study, where the task coordinate team defeat their enemies. To maintain scalable yet effective protocol, introduce Multiagent Bidirectionally-Coordinated Network (BiCNet ['bIknet]) with vectorised extension...
Trust region methods rigorously enabled reinforcement learning (RL) agents to learn monotonically improving policies, leading superior performance on a variety of tasks. Unfortunately, when it comes multi-agent (MARL), the property monotonic improvement may not simply apply; this is because agents, even in cooperative games, could have conflicting directions policy updates. As result, achieving guaranteed joint where each agent acts individually remains an open challenge. In paper, we extend...
Securing coordination between AI agent and teammates (human players or agents) in contexts involving unfamiliar humans continues to pose a significant challenge Zero-Shot Coordination. The issue of cooperative incompatibility becomes particularly prominent when an is unsuccessful synchronizing with certain previously unknown partners. Traditional algorithms have aimed collaborate partners by optimizing fixed objectives within population, fostering diversity strategies behaviors. However,...
Adaptive teaming, the ability to collaborate with unseen teammates without prior coordination, remains an underexplored challenge in multi-robot collaboration. This paper focuses on adaptive teaming multi-drone cooperative pursuit, a critical task real-world applications such as border surveillance, search-and-rescue, and counter-terrorism. We first define formalize \textbf{A}daptive Teaming \textbf{M}ulti-\textbf{D}rone \textbf{P}ursuit (AT-MDP) problem introduce AT-MDP framework,...
Humanoid robots have shown success in locomotion and manipulation. Despite these basic abilities, humanoids are still required to quickly understand human instructions react based on interaction signals become valuable assistants daily life. Unfortunately, most existing works only focus multi-stage interactions, treating each task separately, neglecting real-time feedback. In this work, we aim empower humanoid with reaction abilities achieve various tasks, allowing interrupt at any time,...
We conduct an empirical study on discovering the ordered collective dynamics obtained by a population of intelligence agents, driven million-agent reinforcement learning. Our intention is to put intelligent agents into simulated natural context and verify if principles developed in real world could also be used understanding artificially-created population. To achieve this, we simulate large-scale predator-prey world, where laws are designed only findings or logical equivalence that have...
Abstract With the rapid development of big data science, research paradigm in field geosciences has also begun to shift data‐driven scientific discovery. Researchers need read a huge amount literature locate, extract and aggregate relevant results that are published stored PDF format for building database support In this paper, based on findings study about how geoscientists annotate data, we proposed GeoDeepShovel, publicly available AI‐assisted extraction system their needs. GeoDeepShovel...
The pervasive uncertainty and dynamic nature of real-world environments present significant challenges for the widespread implementation machine-driven Intelligent Decision-Making (IDM) systems. Consequently, IDM should possess ability to continuously acquire new skills effectively generalize across a broad range applications. advancement Artificial General Intelligence (AGI) that transcends task application boundaries is critical enhancing IDM. Recent studies have extensively investigated...
Measuring and promoting policy diversity is critical for solving games with strong non-transitive dynamics where strategic cycles exist, there no consistent winner (e.g., Rock-Paper-Scissors). With that in mind, maintaining a pool of diverse policies via open-ended learning an attractive solution, which can generate auto-curricula to avoid being exploited. However, conventional algorithms, are widely accepted definitions diversity, making it hard construct evaluate the policies. In this...
Trust region methods are widely applied in single-agent reinforcement learning problems due to their monotonic performance-improvement guarantee at every iteration. Nonetheless, when multi-agent settings, the of trust no longer holds because an agent's payoff is also affected by other agents' adaptive behaviors. To tackle this problem, we conduct a game-theoretical analysis policy space, and propose method (MATRL), which enables optimization for learning. Specifically, MATRL finds stable...
Quadruped robots have strong adaptability to extreme environments but may also experience faults. Once these faults occur, must be repaired before returning the task, reducing their practical feasibility. One prevalent concern among is actuator degradation, stemming from factors like device aging or unexpected operational events. Traditionally, addressing this problem has relied heavily on intricate fault-tolerant design, which demands deep domain expertise developers and lacks...
In this work, we investigate the potential of large language models (LLMs) based agents to automate data science tasks, with goal comprehending task requirements, then building and training best-fit machine learning models. Despite their widespread success, existing LLM are hindered by generating unreasonable experiment plans within scenario. To end, present DS-Agent, a novel automatic framework that harnesses agent case-based reasoning (CBR). development stage, DS-Agent follows CBR...
Theory of Mind (ToM) significantly impacts human collaboration and communication as a crucial capability to understand others. When AI agents with ToM collaborate humans, Mutual (MToM) arises in such human-AI teams (HATs). The MToM process, which involves interactive ToM-based strategy adjustment, affects the team's performance process. To explore we conducted mixed-design experiment using large language model-driven agent modules real-time shared-workspace task. We find that agent's does...
Zero-shot coordination (ZSC) is a significant challenge in multi-agent collaboration, aiming to develop agents that can coordinate with unseen partners they have not encountered before. Recent cutting-edge ZSC methods primarily focused on two-player video games such as OverCooked!2 and Hanabi. In this paper, we extend the scope of research multi-drone cooperative pursuit scenario, exploring how construct drone agent capable coordinating multiple capture evaders. We propose novel Hypergraphic...
Yang Li, Cheng Yu, Guangzhi Sun, Hua Jiang, Fanglei Weiqin Zu, Ying Wen, Yang, Jun Wang. Proceedings of the 60th Annual Meeting Association for Computational Linguistics (Volume 1: Long Papers). 2022.
Image distortion during wireless transmission presents a significant challenge for real-world artificial intelligence (AI) applications. Recent methods have attempted to address this issue by integrating neural networks into the system. However, these approaches often require large volume of labeled training data, which can be expensive and time-consuming collect. To issue, we propose novel approach, <bold xmlns:mml="http://www.w3.org/1998/Math/MathML"...
Recent advancements in offline reinforcement learning (RL) have facilitated the training of powerful agents using fixed datasets exclusively. Despite this, quality a dataset plays critical role determining an agent's performance, and high-quality are often scarce. This scarcity necessitates enhancement through subsequent environmental interactions. Particularly, state-action distribution shift may exert potentially detrimental effect on well-initialized policies, thus impeding...